Density Considerations
Thermodynamic Properties
Diffusivity/Viscosity Considerations
Transport Properties
Biological Properties
Solvating Strength
General Process Concepts
Modified Supercritical Fluids

The temperatures normally employed for analytical SFE fall in the 30-150°C range. The materials to be used as SFs must therefore have critical temperatures in this range. The most suitable fluids meeting such a criterion are given in Table 9 in order of increasing critical temperature. Carbon dioxide is the most suitable of the gases listed for analytical purposes since it

• Is relatively nontoxic

• Does not support combustion

• Exhibits readily attainable critical parameters

• Is commercially available in high purity

• Is environmentally compatible

Table 9 Tc and P, of Fluids Used for SFE


Chlorofluorocarbon fluids are implicated in the destruction of the ozone layer in the earth's atmosphere. Nitrous oxide (N2O) will support combustion and tends to decompose spontaneously under certain conditions. Propane is suitable under certain conditions, although it is flammable and forms explosive mixtures with air. The cost of xenon is prohibitive. Water has critical parameters that are not compatible with analytical extraction equipment. Highly polar fluids like NH3, HC1, SO2, and BrF3 are too reactive and toxic, while the solvating power of SF6 is insufficient for all aromatic and polar analytes. In summary, there appears to be no material with critical parameters as mild as those for CO 2 with solvating power as high as CO 2. It is interesting to note that fluoroform (trifluoromethane) has critical parameters lower than CO 2 yet exhibits a superior solvating power to CO 2 for some analytes. The compound CHF3 has a dipole moment of 1.6 D, which is similar to the dipole moment of methanol. Currently, it is not thought to be an ozone depletor; however, its cost may be prohibitive and it is not readily available. Consequently, for the immediate future the advancement of SF tech­nology in the analytical laboratory will be made with CO 2-based fluids. Despite concerns, the use of SF CO 2 is unlikely to contribute significantly to the increase of CO 2 in the atmosphere since a single automobile, for example, will exhaust more than 200 g of CO 2 per mile of travel. Furthermore, the CO 2 used will origi­nate from other sources. However, CO 2 is not ideal in all aspects. For example, mixtures of CO 2 and H2O are corrosive, and CO 2 is not inert with respect to pri­mary and secondary aliphatic amine solutes because both react with CO 2 to yield carbamates. Finally, highly pure CO 2 is not very economical (especially outside the United States), but it is comparable in price with highly pure organic sol­vents.

This practicality of CO 2 use can be illustrated further by considering the sol­vating power of several of these fluids. One measure of fluid solvating power can be estimated from experimentally determined solvatochromic shift data. A com-

parison of the effective solvent polarity of seven fluids as a function of reduced density is shown in Figure 9.29 In this study solvent polarity was defined in terms of solvent polarizability {%*) which is based upon the solvatochromic effect of the fluid on the tt—tc* electronic transition of a probe solute (e.g., 2-nitroanisole) relative to hexane. In this approach mathematical terms are included to account for the effects of polarity and polarizability, but it* does not include effects from potential hydrogen-bonding interaction. At equal reduced densities, the various fluids examined have quite different n* values, indicating large differences in their effective polarity/polarizability. Not surprisingly, NH3 is the most polar flu­id and exhibits the highest solvent power. Xenon shows poorer solvent strength than CO 2 even though xenon has a much higher critical density. Both CO 2 and N 2O show similar solvent strength over the entire density regime even though the latter is a more polar molecule. It should also be noted that changes in n* are most pronounced with changes in density for the more polar fluids.

Fig. 9. The n* solvent polarizability/polarity parameter for various SFs as a function of reduced den­sity (p/pc) at a reduced temperature of 1.03. The it* values are given for comparison.

The 7t* values for liquid methylene chloride, benzene, ethanol, carbon tetrachloride, hexane, and perfluorohexane are also included in Figure 9 for comparison. At liquid-like densities, CO 2 exhibits solvating properties similar to hexane rather than CH 2C 12, as was commonly and erroneously reported in the 1980s.

In a related study, shifts in the absorption wavelength of phenol blue in SF CO 2, C 2H 4, CHF 3, and CF 3C1 were measured (Fig. 10). At a given reduced density and reduced temperature, the solvent strength transition energies (ET's) were similar for CO 2 and C 2H 4. The solvent strength for CF 3C1 was a little weaker than for either CO 2 or C 2H 4, probably because CF 3C1 has a lower critical density. Phenol blue is quite sensitive to any effects of hydrogen bonding, which may explain why SF CHF 3 exhibits the highest solvent power with the phenol blue probe. The ET value of CHF 3, which has an acidic hydro­gen, is comparable to liquid n-hexane at a density of 0.63g/mL, but may be ad­justed to equal that of liquid CC1 4 at a density of 1.19 g/mL.

Fig. 10. The ET (transition energy) of phenol blue as an indicator of solvent strength for various su­percritical fluids (O: C2H4, 25°C; A: CO 2, 45°C; □: CHF 3> 40°C; •: CF 3C1, 30°C).

Designed and Programmed ELECSUS